286 HANDBOOK OF THE OF THE WORLD

A new species of tyrannulet from the upper Madeira-Tapajós interfl uvium in central Amazonian Brazil: Birds don’t always occur where they “should”

Bret M. Whitney1, Fabio Schunck2, Marco Antonio Rêgo2, and Luís Fábio Silveira2

On 8 August 2009, while guiding a Field Guides tour group in undisturbed campina habitat on the left bank of the Rio Madeirinha near the “Pousada Rio Roosevelt” in southern Amazonas state, Brazil, BMW’s attention was drawn to a faint but distinctive, two-note call of unknown identity. Playback of a recording of the vocalization led to fi ne views of aZimmerius 3 tyrannulet that was clearly a vocally distinctive, close relative of the recently described Z. villarejoi (; Álvarez and Whitney 2001) in northeastern Ama- zonian Peru. A good recording of calls of the tyrannulet was made that day, and in September, 2011, Mario Cohn-Haft recorded calls of and collected a fi ne specimen along the BR-230 (“Transamazônica”) highway west of the town of Apuí, Amazonas that he located with playback of BMW’s recording, but it was not until December of that year that a proper fi eld expedition could be mounted to secure a series of voice-recorded specimens to permit a well-documented description. On that expedition, BMW and FS were able to adequately record the vocal repertoire, learn more about the habitat and foraging behavior of the , make photographs and some high-defi nition video of several individuals, and gain a better knowledge of the new species’ distribution. We are now pleased to introduce it as:

Zimmerius chicomendesi Chico’s Tyrannulet Poiaeiro-de-chicomendes (Portuguese)

Holotype.— Museu de Zoologia da Universidade de São Paulo about 60 km east of the Rio Madeira thence east possibly only (MZUSP) 92429, adult male from Brazil: Amazonas; Munici- to the left bank of the Aripuanã/Roosevelt/Madeirinha system, in pality of Humaitá on the BR-230 (“Transamazônica”) highway the state of Amazonas; northern and southern range limits more (07°47'22"S/62°23'32"W), at about 70 m elevation; collected 8 poorly understood, but not known north of about 7°40' south lati- December 2011 by Fabio Schunck, prepared by Marcelo Félix. tude about 45 km west of the mouth of the Rio Roosevelt or south Voice recorded by Bret M. Whitney, original numbers BMW of about 8°40' on the left bank of the Rio Madeirinha (Fig. 1). 14795-97; Macaulay Library of Natural Sounds (ML) 169981. Presently documented only from Amazonas, but we expect that Pectoral muscle tissue preserved in ethanol (MZUSP 92429), it will eventually be found in northeastern Rondônia and north- fi eld number FST-07. western Mato Grosso, perhaps only west of the Rio Madeirinha. It is important to note that the species apparently does not occupy Diagnosis: Morphology.— Based on the series at hand, Zim- some enclaves of suitable habitat near confi rmed points of occur- merius chicomendesi appears to be indistinguishable in plum- rence (see “Range delimitation and Conservation,” below). age or colors of soft parts from Zimmerius villarejoi or the dis- junct “Moyobamba population” probably closely related to that Description of holotype.— See color illustration and Figure 2. Several species (Álvarez and Whitney 2001; hereafter referred to as Z. photos (FS and BMW) and some high-defi nition video (BMW) of the holotype aff. villarejoi). It is notably smaller than these two and females and its mate were made at the time of collection and are archived at the MZUSP. have signifi cantly shorter wing and culmen (bothP < 0.05), and Capitalized color designations (corresponding number in parentheses) from Smithe lighter mass (P < 0.02) than females of Z. villarejoi; there is (1975). Plumage fresh and unworn, no wing or tail molt, tail complete; skull 30% almost no morphometric overlap with Z. aff. villarejoi (see SI). ossifi ed. Upperparts from base of bill to uppertail coverts uniform olive, individual Specimens of all these forms, including the entire series of Z. feathers with Greenish Olive (color 49) bases and Olive Green (S47) tips. Facial villarejoi and four of the fi ve Z. aff. villarejoi, were examined region, throat, breast and sides slightly paler greenish (weak contrast with up- together at the LSUMNS by BMW in March 2012. Voice.— All perparts) washed yellow (throat weakly grayish yellow), becoming clear yellow vocalizations are immediately and primarily distinguishable (brighter than Sulphur Yellow 157) on the belly and undertail coverts. Wings dull from homologous ones of Z. villarejoi and Z. aff. villarejoi in the blackish with a faint olive cast, lesser upperwing coverts with Olive Green (S47) fi eld and in spectrographic analysis by their signifi cantly lower fringes, median and greater upperwing coverts with well-defi ned and more contrast- frequency, among other characteristics (described below). ing yellow (near Sulphur Yellow 157) margins to distal webs (barely extending to proximal web near rachis on some feathers). Primaries 8 and 7 are of equal length Distribution.— Restricted to central Amazonian Brazil from and longest, perhaps just fractionally longer than P9, followed by P6>P5. Three

1 Louisiana State University, Department of Biological Sciences, Museum of Natural Science, 119 Foster Hall, Baton Rouge, Louisiana 70803, USA. ([email protected]) 2 Seção de Aves, Museu de Zoologia da Universidade de São Paulo (MZUSP), Avenida Nazaré 481, Ipiranga, São Paulo, SP, Brazil CEP 04263-000. 3 Zimmerius 9: 295. ORIGINAL SCIENTIFIC DESCRIPTIONS 287

Figure 1. (just before specimen preparation) 100 mm; bill (culmen from base at skull) 8.3 Geographic distribution of Zimmerius chicomendesi in mm; bill from anterior edge of nares 3.7 mm; bill width at anterior edge of nares 2.6 south-central Amazonian Brazil. mm; wing (chord) 43.8 mm; tail 39.6 mm; tarsus 13.3 mm; mass 5.5 g. A red star marks the type locality and letters adjacent Etymology.— Francisco “Chico” Alves Mendes Filho (1944- to red locality dots provide 1988) was a man wise beyond the borders of his time and space. documentation: S = specimen; He learned to use the Amazon rainforest by living in it and un- V = vocal recording. A white derstood the fundamental importance of preservation of natural X marks places searched resources as well as the dire socio-economic consequences of for Z. chicomendesi where their destruction – and Chico Mendes, through perseverance we are reasonably to very confident it is absent. A yellow and, we can only guess (and admire), excellent “politicing,” was outline encompasses points able to make that knowledge count in the international arena. of confirmed occurrence We have no doubt that Mendes and his message, during the last within a significantly more few years of his short life, did more to educate such agencies as extensive region that has not the Interamerican Development Bank to more wisely distribute been inventoried where we funding toward sustainable uses in Amazonia than has any other expect the species is present individual. If Mendes were alive today, we cannot help but im- in appropriate habitat. The agine that Brazil would be far ahead of where it is in the devel- “BX-044 polygon,” a natural opment of a truly sustainable Amazonia in reasonable harmony enclave of mostly cerrado-like habitats, is labeled on the with indigenous peoples and colonists. In bringing this obscure image. Open areas including little bird to the light of science, we call up the spirit of Chico campinas appear pinkish due Mendes to help us all get it right. primarily to high reflectance at mid-infrared wavelengths, REMARKS with recently cleared areas such as those across much of Type series.— The following twelve specimens are the paratypes Rondônia appearing paler pink of Zimmerius chicomendesi: MZUSP 92419-92424 female, fe- than undisturbed habitats. Black male, male, female, female, male, respectively (AM; BR-230 lines mark the boundaries of Brazilian states as indicated at 07°44'15"S/61°05"20"W); and 92427 female, 92428 male, by their official abbreviations: and 92430 female (AM; BR-230 at 07°47'22"S/62°23'32"W); AM = Amazonas; RO = Louisiana State University Museum of Natural Science Rondônia; MT = Mato Grosso. (LSUMNS) 182843 female and 182844 male (AM; BR-230 at The federal highway BR-230 07°47'22"S/62°23'32"W); Instituto Nacional de Pesquisas da (“Transamazônica”) is shown Amazônia (INPA) 2232 male (AM; rodovia BR-230 ca. 145 km in white. wsw Apuí at “campo do coca-cola” 07°43'S/61°04'W). A skel- longest primaries are 4.7 mm longer than the shortest pair (P10 = P4 and three in- eton (“schmoo”) was prepared from MZUSP 92419 and spread ner primaries). Outer webs of primaries and, especially, secondaries very narrowly wings were prepared from this specimen and MZUSP 92420. fringed same yellow as edging on upperwing coverts and contrasting inside the Within the series, there is no significant variation from the dark primary stack; primary coverts unmarked. Secondaries and inner primaries, holotype in plumage coloration or markings, but there is strik- especially on distal portions, similarly but more narrowly margined yellow, lack- ing, sometimes shocking, sexual dimorphism in size, with males ing any expansion of this color near tip of secondaries. Underwing coverts clear, about 5-15% larger in wing and tail measurements (see SI). pale yellow becoming slightly duskier at the wrist. Tail same as wing but rectrices Three mated pairs were collected. Figure 3 shows one of them narrowly fringed with Olive Green (S47), only weakly contrasting with upperparts. (male is MZUSP 92428, female 92427) perched on the same Soft parts in life: iris white, maxilla brownish red, especially along tomia, mandi- branch and about the same distance from the camera lens. Tail ble reddish pink, gape reddish, legs blackish. Standard measurements: total length length and wingspan of the male were 23% and 13% (155 vs. 135

Figure 2. Zimmerius chicomendesi in habitat. A) Birds responded aggressively to playback of conspecific vocalizations by approaching the speaker and cocking the tail; note the white (previously regurgitated) A B mistletoe berry. B) One typical, undisturbed posture. C) The underparts may appear vaguely streaked depending on the lighting and arrangement of the feathers (see also Fig. 3). D) Foraging on the mistletoe Oryctanthus alveolatus. All images by Fabio Schunck (cameras) and Bret Whitney (playback), December, 2011, along the BR-230 (“Transamazônica”) highway. High-definition videos of some of these encounters are available for viewing on the C D IBC website. 288 HANDBOOK OF THE BIRDS OF THE WORLD

Zimmerius chicomendesi forages regularly, perhaps mostly, Figure 3. A mated pair of Zimmerius on fruits of the widespread mistletoe (Loranthaceae) Oryctan- chicomendesi. The sexes thus alveolatus (Kunth) Kuijt and regurgitates the sticky seeds, are often, but not always, of which are wiped on twigs and branches of trees and shrubs, ef- remarkably different size, fectively dispersing the plants. This is the same plant identified males (upper bird) larger than as an important food source of Z. villarejoi, and foraging be- females. See SI for a photo havior is the same as described for that species (Álvarez and of this pair after specimen Whitney 2001) except that we did not observe Z. chicomendesi preparation. Image by Fabio perform a single foraging maneuver obviously directed at cap- Schunck (cameras) and Bret ture of arthropod prey. Examination of stomach contents of 8 Whitney (playback), December, 2011, along the BR-230 individuals revealed only fruit-pulp residue and a few seeds (“Transamazônica”) highway. that appeared to belong to the mistletoe identified above. A few minute insect fragments may have been ingested incidentally with mistletoe berries, but the bird must also eat some arthro- pods, the diet perhaps varying seasonally. Two other small, mistletoe-specialist tyrannids are sympat- ric with Zimmerius chicomendesi: Z. acer () and Tyrannulus elatus (Yellow-crowned Tyrannulet). Both of these species are larger than Z. chicomendesi, sex for sex, and a Z. acer pair recorded and collected 11 December 2011 (MZUSP 92406 male and 92407 female) showed size dimorphism similar to that described above for Z. chicomendesi. Both Z. acer and T. elatus prefer taller trees and forest-edge such as that found around the borders of campinas. Recording playback of Z. chicomendesi at the ecotone of campina with tall forest brought the above-mentioned pair of Z. acer immediately down into the campina where one of them rapidly chased off a Z. chicomende- si. Further playback of chicomendesi caused the acer to stay in the edge of the campina, guarding two large clumps of Oryctan- thus misletoe; it was clearly dominant over Z. chicomendesi but meetings between these two must be limited to the ecotone, as neither significantly infiltrates the realm of the other. A pair of Tyrannulus elatus well out into the campina but keeping to a copse of taller trees also displayed dominance over neighboring mm) greater than those of his mate, and mass was 21% heavier. Z. chicomendesi after being agitated with whistled imitations of To our knowledge, this level of size dimorphism has not been its song. It seems plausible, even probable, that interspecific in- reported in tyrannids having no specialized feather structures teractions of Z. chicomendesi with these two tyrannulets could (e.g., Yetapa, Alectrurus, Muscivora). None of Z. villarejoi, Z. have a limiting effect on its ability to disperse through taller for- aff. villarejoi, or Z. chicomendesi ever seems to attain complete est separating isolated campinas. skull ossification, and more than 50% ossification is unusual. We found Z. chicomendesi in mated pairs in about half of One specimen of Z. chicomendesi (MZUSP 92423) is an im- our observations in December; we did not note any behavior that mature replacing worn, apparently juvenal plumage (rectrices suggested nesting and gonads were not well developed at this noticeably narrower and more pointed than those of adults) with (early wet) season; BMW’s more limited observations in much fresh adult feathering. drier August did not suggest breeding at that time, either. On three occasions (recorded once in response to recording play- Ecology and behavior.— Zimmerius chicomendesi inhabits back of the song, BMW-14932), males were observed to pro- only campina woodland and scrub growing on flat, sandy, of- duce a single brief, quiet, mechanical buzz with the wings while ten poorly drained ground (Fig. 4A), or on compacted, rock- flying between perches in neighboring trees. No unusual flight like sand in well-drained, hillier and grassier terrain (Fig. 4B). trajectory or other posturing during production of the mechani- Structure of woody vegetation is similar on these rather differ- cal sound was discernible. High-definition video of Zimmerius ent substrates, dense and interlocking around irregular openings chicomendesi in habitat may be viewed on the Internet Bird Col- almost devoid of vegetation; average height varies from about lection (IBC) website. 2 m around openings to roughly 6 m in the densest areas, with scattered, mostly Caraipa spp. (Clusiaceae) trees reaching about Vocalizations.— Of twelve specimens collected in December, 10 m in height. 2011, eleven were recorded first. In addition, we have calls and Figure 4. Two campina biotopes inhabited A B by Zimmerius chicomendesi. A) Flat, loosely sandy (quartzitic) ground along the BR-230 (“Transamazônica”) highway. Fragile campina vegetation may never recover from damage caused by roads. Image by Fabio Schunck December, 2011 (wet season). B) Undisturbed campina growing on well- drained, compacted, rock-like sand at the site of the species’ discovery on the left bank of the Rio Madeirinha in southern Amazonas. Zimmerius acer and Tyrannulus elatus are present in the taller forest in the background. Image by Bret Whitney August, 2009 (dry season). ORIGINAL SCIENTIFIC DESCRIPTIONS 289

Figure 5. 7 Vocalizations of Zimmerius A B C D chicomendesi in south-central 6 Amazonian Brazil in comparison 5 to homologous vocalizations of its presumed closest relatives, 4 in Peru. (A) One typical song of Z. chicomendesi (BMW-14976, 3 holotype MZUSP 92430); (B) 2 Infrequently heard variant of song of Z. chicomendesi 1 featuring distinctive introductory note (BMW-14932, LSUMNS 0 182844, 10 December 2011); 0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 (C) One typical song of Z. 7 villarejoi (Peru: Loreto; about E F G 6 24 km west of Iquitos at (kHz) Frequency Allpahuayo-Mishana Reserve, 5 1 Aug 1997; Whitney); (D) One song of Z. [villarejoi] (Peru: San 4 Martín; Quebrada Upaquihua se 3 of Tarapoto, July 2003, Daniel F. Lane); (E) Two typical two-note 2 calls of Z. chicomendesi (Brazil: Amazonas; left bank of Rio 1 Madeirinha 08 August 2009; 0 Whitney); (F) Typical two-note 0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 call of Z. villarejoi (Peru: Loreto; Time (seconds) about 24 km west of Iquitos at Allpahuayo-Mishana Reserve, 1 Aug 1997; Whitney); (G) Two songs of at least four individuals recorded well by BMW on the phologically divergent population will just as certainly fit into variants of two-note calls from a single individual of Z. [villarejoi] left bank of the Rio Madeirinha above “Pousada Rio Roosevelt” the above clade but as a separate species, probably as sister to (Peru: San Martín; Quebrada in August 2009 and in December 2010, 2011, and 2012. We Z. villarejoi; its description will be published elsewhere (Lane Upaquihua se of Tarapoto, July documented four types of vocalizations of Zimmerius chico- et al. in prep.). 2003, Daniel F. Lane). mendesi: song and common call (both heard and recorded regu- Figure 6 shows the distributions of the three members of larly); call series (heard once in the field and recorded once from the postulated Zimmerius villarejoi complex. Other, undiscov- a bird in hand), and a snarl (quiet but aggressive sound emitted ered members should now be sought in unexplored regions of after recording playback and probably used in natural, aggres- the Amazon basin, perhaps especially south of the Marañón/ sive interactions). This repertoire parallels those documented for Solimões rivers in Peru and Brazil. Álvarez and Whitney (2001) Z. villarejoi (Álvarez and Whitney 2001) and Z. aff. villarejoi considered all of the “red-billed” Zimmerius tyrannulets, includ- (Daniel F. Lane recordings), including basic syntax of most vo- ing the significantly larger and vocally further-differentiated Z. calizations, but all sounds are notably lower in frequency (see cinereicapillus (Red-billed Tyrannulet), to form a natural group Fig. 5 for inter-taxon comparisons, and it is recommended that more closely related to each other than to any other species. We interested readers also listen to the sounds on the IBC website; agree that Z. villarejoi, in a complex including the two more re- an audio comparison is "worth a thousand words"). Both sexes cently identified taxa above, will prove to be sister toZ. cinerei- emit all of the above vocalizations, and their voices are essen- capillus, which seems to be patchily distributed in the foothills tially indistinguishable. The most frequently heard vocalization of the east slope of the Andes from central Ecuador south to is a distinctly two-noted call, with one- and three-note calls de- northern Bolivia. This relationship has been partially validated livered less regularly. Individuals are capable of tweaking these very recently by the finding of Rheindt et al. (2013) that Z. cin- notes to make them shorter or longer, or add inflections, which ereicapillus is sister to Z. aff. villarejoi, which they considered we suppose could function to relay different kinds of informa- conspecific with Z. villarejoi. The work of Rheindt et al. (2013) tion. Unfortunately, we were not able to record the dawn song provided significantly better resolution of the evolutionary his- of Z. chicomendesi, but based on a vocalization heard early one tory of the genus Zimmerius than that presented in Rheindt et al. drizzly morning by BMW before he could get the microphone (2008). Our knowledge of diversification in the genus will take ready, it may be a single note like one of the notes in the normal another leap forward when more outstanding taxa, such as Z. call repeated at intervals of about 3-5 seconds. villarejoi and Z. chicomendesi, are included. Furthermore, a ma- jor expansion of the Zimmerius radiation to include the Atlantic Relationships and .— When a taxon-complete, DNA- Forest biome of eastern Brazil will likely result from inclusion based phylogenetic analysis is eventually undertaken for the of Phyllomyias griseocapilla (Gray-capped Tyrannulet) in the genus Zimmerius, Z. chicomendesi will surely prove to be in a phylogeny (BMW pers. obs.). well-defined clade with the recently describedZ. villarejoi (Mis- hana Tyrannulet, Álvarez and Whitney 2001), another rare bird Range delimitation and Conservation.— That Zimmerius chi­co­ of highly local distribution in varillales (as campinas described men­desi is among the most locally distributed (among known!) above are called in northern Peru) near Iquitos, Peru. For the species in all of Amazonia and requires a rare habitat type with- description of Z. villarejoi, the single specimen from near Moy- in its range raises immediate concern despite that fact that it is obamba, Departamento San Martín (Field Museum of Natural common where it does occur. The species is, however, difficult History [FMNH] 49405), was examined but not included as to find if its voice is not recognized, and it was not detected a paratype because it was taken quite some distance from the during previous visits to the site of discovery (e. g., Whittaker type locality of villarejoi, its habitat was not recorded, and its 2009). Broadcast of recordings of the calls and song in appro- voice was unknown. In July, 2003, Daniel F. Lane located this priate habitat in both dry and wet seasons elicits a response Moyobamba population (in this paper called Zimmerius aff. vil- within 1-20 minutes (almost always less than 5 min) if the spe- larejoi) at “Quebrada Upaquihua” about 26 km south-southeast cies is present. of Tarapoto and tape-recorded and collected five males; subse- For conservation purposes, Zimmerius chicomendesi is quently, this form has been found throughout the Río Mayo val- appropriately viewed as an “interfluvial campina specialist ley and along the middle Río Huallaga, in drier and white-sand [ICS],” an idea easily understood when campinas in the Ma- habitats. It was considered a disjunct, conspecific population of deira-Aripuanã interfluvium, or other Amazonian interfluvia, are Z. villarejoi by Schulenberg et al. (2007). This vocally and mor- identified in satellite images as scattered and often small patches 290 HANDBOOK OF THE BIRDS OF THE WORLD

Figure 6. Distribution of the postulated Zimmerius villarejoi complex in South America. White dots are Z. villarejoi in Loreto, Peru (white star marks the type locality at the Allpahuyo-Mishana Reserve outside Iquitos); black dots are Z. aff. villarejoi in San Martín, Peru, referred to as “Moyobamba population” in Álvarez and Whitney (2001); and red dots are Z. chicomendesi in Amazonas, Brazil (red star marks the type locality along the BR-230 [“Transamazônica”] highway). Additional close relatives, as yet undiscovered, should be sought in campinas, perhaps especially in the vast, largely unexplored region south of the Marañón/Solimões rivers in Peru and northwestern Amazonian Brazil.

(Fig. 1, viewable in much greater detail online in SI). We used presence. The yellow outline in Figure 1 on the right bank of examination of satellite imagery to mount our highly success- the Rio Madeira encompasses 69,706 km2 within the Aripuanã- ful December, 2011, expedition to collect a series of the new Madeirinha-Machado interfluvium. Our working hypothesis, Zimmerius based on expectation that easily accessible campinas after determining absence of Z. chicomendesi in suitable habitat identifiable along the BR-230 (through which BMW had hitch- across the Madeirinha and Roosevelt rivers and in recognition hiked several times without stopping) would likely produce it. of the emerging pattern of endemism in this "speciation block" We had employed the same reasoning, however, in a previous or "mini-interfluve" (Cohn-Haft et al. 2007), is that Z. chico- (August, 2011), failed attempt to find the bird in an extensive mendesi occurs only or almost entirely within this area. Under campina that appeared to be appropriate in satellite imagery on IUCN (2001) definition, this is the species’ estimated “extent the left bank of the Rio Roosevelt about 60 km southeast of the of occurrence”. The approximate total area of what appears site of discovery of the species and, most recently (December, in satellite imagery to be potentially suitable campina habitat 2012), at a campina only 26 km distant on the right bank of the within this outline is 5442 km2. Of this, a large enclave of rela- Roosevelt. tively open habitats known as the BX-044 polygon (Aleixo and Assuming that our search of what we are confident was ideal Poletto 2007) accounts for about 2500 km2. We are confident habitat at all of these sites and more cursory inspection of an- that Z. chicomendesi almost certainly does not occupy the great other one about 80 km northwest of the site of original discov- majority of the BX-044, which is dominated by inappropriate ery but of somewhat different plant composition and structure grasslands, cerrado-like plant associations, and gallery forests would have revealed the presence of the bird if it occurs at these (Aleixo and Poletto 2007 and ground-truthed by BMW and places, this indicates that Z. chicomendesi does not occupy all FS December, 2011), with campina/campinarana distributed potentially suitable habitat quite near to known points of occur- mainly around parts of the periphery. Notwithstanding that few rence. Thus, our “educated modeling” exercise – visual inspec- areas of the BX-044 have been adequately searched for the new tion of satellite imagery to identify suitable campinas followed species, the only records of it are those from the extreme south- up by ground-truthing to verify habitat suitability – proved inad- eastern tip. equate to predict either occurrence or absence of the new spe- We drew a line 1 kilometer wide around the BX-044 to ap- cies and indicates that other, paleohistorical and geographical proximate a belt of campina habitat that was, in available satel- factors, probably in combination with dispersal capability and lite imagery, difficult to discern, in an attempt to estimate the competition from ecological correlates, have had deterministic area of habitat potentially appropriate for the bird. This was 630 effects in restricting its contemporary distribution. km2. Thus, we arrived at a total of 3572 km2 of habitat within This case highlights the critical importance of ground-truth- our yellow outline that we expect is the “area of occupancy” ing specific sites evaluated as being appropriate for organisms (IUCN 2001) of Zimmerius chicomendesi. We estimate that the to document actual presence or, just as importantly, absence population density averages about three pairs/km2, which results there, especially in the cases of rare and endangered species. Of in a global population estimate of 10,700 pairs. Whether or not course, the validity of our exercise is based on how accurately or this calculation proves to be nearly accurate, we do expect that completely we evaluated habitat suitability and also absence of the present population size of Z. chicomendesi is not much dif- the species in places it was expected to occur in and around the ferent from what it was before the Transamazônica was pushed area of known occurrence. We would welcome with great inter- through in the 1970s; in other words, it is almost certainly still est the results of future research, including our own, that might close to its natural equilibrium. show that our evaluations were erroneous; for the time being, Some 50% of the postulated range of Zimmerius chico- the burden of proof lies there. We looked for Z. chicomendesi mendesi lies within several indigenous reserves, the recently or relatives of it in four additional campinas more remote from created Campos Amazônicos National Park, and some other known points of occurrence, four east of the Rio Madeira south smaller federal and state conservation areas, but we can antici- of Jaci-Paraná, Rondônia and one on the left bank of the Rio pate specific threats that must be addressed. The most serious Sucunduri (5°48’S/59°16”W), but it was apparently absent in of these centers on consequences of the Brazilian government all of these places. Fieldwork in campinas west of the Rio Ma- eventually deciding to pave (or try to pave) the BR-230, which deira between Porto Velho, Rondônia and Humaitá, Amazonas cuts through much excellent campina habitat in the Madeira- has probably been extensive enough to establish absence of the Aripuanã interfluvium (Fig. 7). As it is (still a dirt road), the species-complex there as well. “Transamazônica” highway completely traverses the range of The survival outlook for Zimmerius chicomendesi current- Z. chicomendesi and raw materials for paving it, mainly sand, ly is not alarming as much of its range is remote from human would be quarried heavily resulting in irreversible damage to ORIGINAL SCIENTIFIC DESCRIPTIONS 291

Figure 7. The federal highway BR-230 (the “Transamazônica”) was constructed in the 1970s. This dirt road traverses the entire Madeira-Tapajós interfluvium, passing through many campinas inhabited by Zimmerius chicomendesi and several other similarly range-restricted birds. Clearing along either side of the highway currently ranges from about 10 meters to more than five kilometers. When the Brazilian government eventually decides to pave the Transamazônica, fragile campina habitats near the road, all currently unprotected except where they are marginally safer in indigenous reserves, will be heavily and probably irreversibly damaged. Now is the time to ensure that this habitat is preserved. Perhaps the most efficient and effective way to do this would be to incorporate these campinas in the nearby Campos Amazônicos National Park, with signs clearly marking the stretches of the highway fragile, incredibly slow-growing campina habitats (Fig. 4A). joi from San Martín, Peru, and provided helpful comments on where no extractive activities or other damage, such as fire or Intricate but vitally important local drainage patterns would be an early draft of the paper. José "Pepe" Álvarez kindly sent us homesteading, are permitted. interrupted haphazardly with dire consequences for campina his list of confirmed localities for Z. villarejoi in Loreto, Peru. Image by Fabio Schunck, plant communities. And because campina is not a fire-adapted Mario Cohn-Haft and Ingrid Macedo of INPA provided help- December, 2011. habitat (different from some other low-stature, open-vegetation ful photographs and other information pertinent to the specimen communities), there would most definitely be damaging burns Mario collected in September, 2011. FS is particularly thankful concomitant with roadwork and the inevitable encroachment of to João Bosco and José Luiz Gonçalves for help with armaments human settlement, or simply with increased traffic through this in São Paulo that greatly facilitated the collection of specimens still-remote region. Significant stretches along the BR-230 in the in excellent condition in December, 2011. Botanical colleagues western Madeira-Tapajós interfluvium are somewhat protected Lúcia Lohmann and, especially, Claudenir Caires identified the by indigenous territories, which is fortunate because these native mistletoe species from our photos. Phyllis Isler kindly prepared people do not tend to exploit campinas for resources that result the spectrograms, and Vagner Cavarzere was very helpful in per- in habitat degradation – although it is to be expected that ex- forming the statistical analysis of morphometric data. Richard ploitation would accelerate suddenly as a paved BR-230 would Banks, Lars Pomara, and Thomas Schulenberg provided helpful stimulate expansion of roadside villages, leading to quarrying of comments on the manuscript. Hilary Burn painted the figure of sands for construction. Z. chicomendesi that accompanies this description. Right now, before invasive activities begin, official protec- tion of the campinas bordering the BR-230 could be established Literature Cited with delimitation of them as protected habitats that cannot be Aleixo, A. and F. Poletto (2007). Birds of an open-vegetation enclave exploited for construction material or other resources and must in southern Brazilian Amazonia. Wilson Journal of Ornithology 119: not be settled or burned. Ideally, we suggest, these campinas 610–630. should be incorporated in nearby Campos Amazônicos National Álvarez Alonso, J. and B. M. Whitney (2001). A new Zimmerius tyran- Park and signs marking their status as protected areas of the park nulet (Aves: Tyrannidae) from white sand forests of northern Amazo- should be posted along the highway. Of course, official protec- nian Peru. Wilson Bulletin 113: 1–9. tion of the campinas would not guarantee their effective protec- Cohn-Haft, M., A. M. F. Pacheco, C. L. Bechtoldt, M. F. N. M. Torres, tion, but it would be enormously wise and helpful. A. M. Fernandes, C. H. Sardelli, and I. T. Macêdo (2007). Capítulo 10. Inventário ornitológico. Pp. 145–178 in: Rapp Py-Daniel, L., C. Acknowledgments.— We are especially grateful to Marcelo P. Deus, A. L. Henriques, D. M. Pimpão, and O. M. Ribeiro (orgs.). Félix for excellent specimen preparation and camaraderie in Biodiversidade do Médio Madeira: Bases científicas para propostas the field and to Luis Cleiton Holando Lobato for fine logistical de conservação. INPA: Manaus, 244 pp. assistance and secure travel on precarious Amazonian roads in IUCN (2001). IUCN Red List Categories and Criteria: Version 3.1. the early rainy season of 2011. Many thanks to Waner Costa of IUCN Species Survival Commission. IUCN, Gland, Switzerland and “Pousada Rio Roosevelt” in southern Amazonas state for out- Cambridge, UK, 30 pp. standing service and comfort provided over the course of sev- Rheindt, F. E., A. M. Cuervo, and R. T. Brumfield. (2013). Rampant eral years to Field Guides tour groups under BMW’s guidance. polyphyly indicates cryptic diversity in a clade of Neotropical fly- Thanks to the Fundação de Amparo à Pesquisa no Estado de catchers (Aves: Tyrannidae). Biological Journal of the Linnaean So- São Paulo (FAPESP) and Conselho Nacional de Desenvolvi- ciety 108: 889–900. mento Científico e Tecnológico (CNPq) for the concession of Rheindt, F. E., J. A. Norman, and L. Christidis. (2008). DNA evidence grants (Evolução da Fauna de Vertebrados Terrestres Brasilei- shows vocalizations to be a better indicator of taxonomic limits than ros do Cretáceo ao Presente: Paleontologia e Filogenia, CNPq plumage patterns in Zimmerius tyrant-flycatchers. Molecular Phylo- 565046/2010-1), fellowships (LFS) and for the authorization genetics and Evolution 48: 150–156. for collecting and Research by Foreigners and also to the In- Smithe, F. B. (1975). Naturalist’s color guide. American Museum of stituto Brasileiro do Meio Ambiente e dos Recursos Naturais Natural History, New York. Renováveis (IBAMA – SISBIO) for collecting permits. We Whittaker, A. (2009). Pousada Rio Roosevelt: A provisional avifaunal are grateful to NASA for free and open access to the MODIS inventory in south-western Amazonian Brazil with information on (EOSDIS) satellite imagery used to produce the map image. Dan life history, new distributional data and comments on taxonomy. Cot- Lane generously made available his recordings of Z. aff. villare- inga 31: 20–43.